The Time-Transgressive Termination of the African Humid Period

Home , African humid period, Dahomey Gap, Lake Tana, Paleohydrology, Selima Oasis, Wadi Howar

SUPPLEMENTARY INFORMATION DOI: 10.1038/NGEO2329

The time-transgressive termination of the African Humid TheThe time-transgressive time-transgressive termination termination of the African of Humid the African Period Humid Period Period

Timothy M. Shanahan, Nicholas P. McKay, Konrad A. Hughen, Jonathan T. Overpeck,

TimothyBette Otto-Bliesner, M. Shanahan, Clifford Nicholas W. Heil,P. McKay, John King,Konrad Christopher A. Hughen, A. Jonathan Scholz, John T. Overpeck, Peck

Bette Otto-Bliesner, Clifford W. Heil, John King, Christopher A. Scholz, John Peck Supplementary Information

Supplementary1. Supplementary MethodsInformation

1.1.1. Supplementary Study Area Methods Lake Bosumtwi is a small (~8 km diameter) but deep (~75 meter) stratified lake 1.1. Study Area occupying a meteorite impact crater in the tropical forest zone of southern Ghana (Fig. Lake Bosumtwi is a small (~8 km diameter) but deep (~75 meter) stratified lake S1, S2). The lake is internally draining and isolated from the regional groundwater occupying a meteorite impact crater in the tropical forest zone of southern Ghana (Fig. system, making it exceptionally sensitive to changes in the precipitation-evaporation S1, S2). The lake is internally draining and isolated from the regional groundwater balance 1-3. Sheltering of the lake by the surrounding crater walls limits deep mixing of system, making it exceptionally sensitive to changes in the precipitation-evaporation the water column, resulting in permanently anoxic bottom waters and the preservation of balance 1-3. Sheltering of the lake by the surrounding crater walls limits deep mixing of fine (mm-scale) laminations, which have been previously demonstrated to be annual in the water column, resulting in permanently anoxic bottom waters and the preservation of nature 4. Most of the surrounding catchment is forested, except the flat-lying areas, fine (mm-scale) laminations, which have been previously demonstrated to be annual in which have been converted to agriculture (e.g., maize, plantain, cassava) over the last few nature 4. Most of the surrounding catchment is forested, except the flat-lying areas, decades 5. which have been converted to agriculture (e.g., maize, plantain, cassava) over the last few In 1999 and 2004, a series of freeze cores, piston cores and drill cores were decades 5. collected from the lake depocenter (Fig. S1)6. The freeze cores capture the sediment In 1999 and 2004, a series of freeze cores, piston cores and drill cores were water interface intact and were visually correlated to the piston cores using marker collected from the lake depocenter (Fig. S1)6. The freeze cores capture the sediment laminations that could be traced in sediment cores across the central basin 4. The piston water interface intact and were visually correlated to the piston cores using marker cores are finely (mm-scale) and continuously laminated over their entire length, with the laminations that could be traced in sediment cores across the central basin 4. The piston exception of the interval between 2.1 and 3.9 meters, which is organic-rich and contains cores are finely (mm-scale) and continuously laminated over their entire length, with the abundant well-preserved remains of the cyanobacteria Anabaena indicative of eutrophic exception of the interval between 2.1 and 3.9 meters, which is organic-rich and contains conditions (Unit S1) 4. Drill cores collected during the Intercontinental Drilling Program abundant well-preserved remains of the cyanobacteria Anabaena indicative of eutrophic (ICDP) sediment coring expedition to the lake in 2004 provide a nearly continuous 290- conditions (Unit S1) 4. Drill cores collected during the Intercontinental Drilling Program (ICDP) sediment coring expedition to the lake in 2004 provide a nearly continuous 290-

NATURE GEOSCIENCE | www.nature.com/naturegeoscience 1 1

© 2015 Macmillan Publishers1 Limited. All rights reserved meter long sediment record, which spans the last ca. 1.07 Ma 6. The present study focuses on the uppermost 10.6 meters of this sediment sequence preserved in the piston cores, which covers the last 25 kyr. Located in southern Ghana, the climate of Lake Bosumtwi is heavily influenced by the West African monsoon, which draws moisture onshore from the Gulf of Guinea during the months of April-October as the monsoon trough migrates northward with the sun (Fig. S2). In winter the winds reverse, and the climate of southern Ghana is influenced by hot, dry northeasterly monsoon winds, which inhibit precipitation over much of North Africa. Because of its near equatorial (6°N) location and its proximity to the coast, the rainy season is longer and precipitation totals are high near the coast and around Lake Bosumtwi (e.g., 1260 mm yr-1) and decrease with increasing latitude 3.

1.2 Age depth modeling A radiocarbon chronology for the sediment record is based on Bayesian age-depth modeling of 107 radiocarbon ages on plant macrofossils and bulk organic matter covering the upper 21.4 meters (Fig. S3) 7. Modeling was performed with the R software package BACON 8 and using the IntCal09 radiocarbon calibration curve 9. Radiocarbon dates from overlapping cores were correlated using matches from marker laminae and magnetic susceptibility profiles and placed on a standardized depth scale (relative meter composite depth scale – RMCD)7. For modeling purposes, BACON uses the prior assumption that ages are in stratigraphic order and that sedimentation rates fall within a designated range and follow a long-tailed distribution. The program models sedimentation rates using a simple gamma autoregressive process and generates posterior age distributions using a self-adjusting Markov Chain Monte Carlo algorithm 10. The result is a population of possible age-depth models, which reflect the probability distributions associated with the analytical radiocarbon age uncertainties as well as the uncertainties in the calibrated ages.

1.3 Reconstruction of paleolake highstands The paleolake level history was reconstructed using a combination of radiocarbon dated highstand benches identified above the modern lake level, lowstand terraces

© 2015 Macmillan Publishers Limited. All rights reserved identified in seismic data and sediment cores, and lacustrine muds deposited above the modern lake and preserved on the crater walls (Fig. S4, main text, Fig 2b.; for a more detailed description see 1,11). Radiocarbon dates on roots from a submerged paleoterrace indicate that the lake was at least 60 meters lower at 16.3 ka 1,12. However, laminated lacustrine silts 20 meters above the modern lake level indicate that the lake was already substantially deeper than today by 14.5 ka. This recovery was interrupted by a lake level regression between 12.9-11.2 ka, followed by a return to deep lake conditions in the early Holocene. The early Holocene highstand deposited extensive lacustrine silts on the walls of the crater, up to 110 meters above the modern lake, until at least 5.7±0.4 ka 1, though significant undated lacustrine silt accumulations above this date indicate that the high lake stand persisted much later than this. A wave-cut notch in the crater rim indicates that the positive water balance during the early Holocene resulted in the lake overflowing the crater rim 4. Although the lake level record does not completely constrain the end of the Holocene lake level highstand, a well-dated terrace 15 meters above the modern lake level suggest that it had mostly receded by ca. 2.8-3.5 ka. Submerged terraces, identified in seismic data and confirmed in proxy data suggest that the lake shrank by more than 25 meters below modern lake level at least twice in the last 2 millennia, the most recent of which occurring coincident with the Little Ice Age (1500-1800 AD) 1,13. Constraints on the timing of the maximum lake level come from radiocarbon dating of a maximum highstand terrace ca. 110 meters above the modern lake surface and cosmogenic surface exposure dating of wavecut bedrock in the overflow spillway at the same elevation. Radiocarbon dates on charcoal from the base of the highstand terrace yield an age of 8.6-9.0 ka, which we interpret as indicating the onset of overflowing conditions 1. Cosmogenic radiocarbon surface exposure ages on eroded bedrock from the overflow provide an estimate of when water ceased overflowing from the crater. Earlier exposure ages estimates were similar to that obtained from the highstand terrace (9.5±1.5 ka) 1. However, a revised age estimate using updated cosmogenic 14C production rate estimates and scaling procedures (assuming a spallation production rate of 13.6 atom/gram/yr at sea level and high latitude 14; and computed with the Cronus calculator 15), yielded significantly younger ages (5.9±0.9, 6.2±0.9 ka). Together these data suggest

1.4 Stable isotope analysis of leaf waxes 1.4.1 n-alkane extraction and separation 1-5 gram sediment samples were freeze dried and solvent extracted using either a Dionex Accelerated Solvent Extraction System or a CEM MARS X Microwave Extraction System with dichloromethane:methanol (9:1; v/v) to obtain a total lipid extract (TLE). The TLE was partitioned into acid and neutral fractions by aminopropyl flash chromatography using 9:1 DCM:MeOH and 2% formic acid in DCM. The neutral fraction concentrated under N2 and separated into polar and apolar fractions over silica gel with hexane and MeOH. Either urea adduction or molecular sieve was used to remove branched compounds and when needed, Ag+ chromatography was used to isolate the alkane and alkene fractions prior to stable isotope analysis.

1.4.2 Hydrogen isotope analysis of n-alkanes Hydrogen isotope analysis of individual n-alkanes was performed using an Agilent 6890 gas chromatograph operated in splitless mode and equipped with a DB-5 ms column (30m x 0.25µm x 0.25mm), coupled to a Delta V isotope ratio mass + spectrometer via a pyrolysis interface operated at 1430°C. H3 factors were determined daily, and external isotope standards (either the B2 n-alkane or F8 FAME mix, Indiana University Biogeochemical Laboratories) were measured between every 6 to 8 samples. The external standards had a precision of <±5 ‰. Coinjection of propane during the analysis of the external standards allowed us to determine the isotope value of our propane reference tank and allowed us to use the propane as an internal standard within each unknown sample analysis. Between three and five propane injections were performed during each sample run. Within each run the propane peaks had a precision of <±5 ‰. Most samples were run in either duplicate or triplicate and the standard deviations based on replicate analysis of the same samples was typically <±3 ‰.

© 2015 Macmillan Publishers Limited. All rights reserved 1.4.3 Correcting leaf wax hydrogen isotopes for ice volume and vegetation effects In this study, we interpret changes in δDwax as reflecting changes in the

δDprecipitation, which in turn is associated with changes in rainfall amount. On glacial- interglacial timescales there is an additional effect on the δD of precipitation associated with changes in the isotopic composition of source waters due to changes in ice volume. To correct for this effect, we assumed a 1‰ change in δ18O at the Last Glacial Maximum (LGM; 16,17. We converted this change in δ18O to a change in the δD of the glacial ocean assuming that the deuterium excess of the ocean was zero 18,19. We then developed a time series of δDocean values spanning the last 20,000 years by scaling these changes to 20 the LR04 benthic oxygen isotope stack . Changes in δDocean were then subtracted from measured δDwax values to produce a corrected δDwax record (δDcorr.) Previous studies of modern vegetation have demonstrated that there is a significant negative isotopic fractionation associated with the incorporation of hydrogen 21 into leaf waxes . This offset appears to be significantly different for plants using the C3 21 and C4 photosynthetic pathways , leading to a significant influence on δDwax values when vegetation changes are large (as in this study). To address this, we follow a procedure similar to that used by other workers (e.g., 22) and use the δ13C values of the terrestrial leaf waxes to adjust the δDwax values for the effect of vegetation type. 13 Assuming that the dominant control on the δ Cwax signal is the relative proportions of C3 woody plants and C4 grasses, we estimate that the proportion of C4 grasses shifted from ~60% in the deglacial to ~10% in the late Holocene. The magnitude, abruptness and approximate timing of this change are in good agreement with a previously published, lower resolution record of changes in grass pollen from Lake Bosumtwi23 (% grass pollen shifts from 60% to ~5%), supporting our assumption that the δ13C signal is 13 predominantly the result of changes in vegetation type. Using these δ Cwax-derived estimates of vegetation changes, we estimate the apparent fractionation factor between the δDwax (for the C31 n-alkane) and the δDprecipitation(εwax-precip.) using endmember values reported for grasses (-145±15‰)24 and tropical forests (-125±5‰)25 following previous workers22.

Because of the large change in δ13C between the glacial/deglacial period prior to 14.6 ka, and the Holocene, this adjustment significant alters the relative magnitude of the changes in δDwax over the deglaciation. However, it makes relatively little difference in the overall timing or rate of δD changes in the record (Fig. S5).

1.5 Synthesis of paleohydrologic records

To better understand spatiotemporal variations in North African hydroclimate, we synthesized available published records from the literature (Table S1, Fig S6a,b, Fig. S7 animation, main text, Fig. 3). The synthesis draws upon near-shore paleomarine and paleolimnological records based on a wide variety of proxies including stable isotopes (δ18O, δ 13C), biological indicators (diatoms, ostracodes, molluscs), stratigraphic dating of paleolacustrine deposits and lithological changes, palynological reconstructions of vegetation changes, and changes in geochemistry. We also include all published δDwax records spanning the time interval of interest and records of changing terrigenous dust fluxes in marine records. All reported uncalibrated radiocarbon ages were calibrated using CALIB7.026. Where possible, we used the interpretations of the datasets as reported in the original manuscripts or in existing synthesis papers e.g.,27. Individual site data was plotted using the clockwheel approach used previously for the Asian monsoon system 28. We excluded data from the non-peer reviewed literature, in which paleohydrologic or vegetation inferences were uncertain, and at sites where significant portions of the AHP were not covered by the reconstructions. To simplify the diagram, we reduced the stable isotope, paleohydrologic and the paleovegetation data to three hydrologic phases: high, intermediate (or transitional) and low. We recognize that these designations may be somewhat subjective, and particularly with the pollen data, may be complicated by other factors such as non-analog vegetation assemblages, local (non-hydrologic) controls on vegetation and anthropogenic modification of the landscape. High values are intervals when the inferred hydrologic balance was at or near the early Holocene maximum. Low values are when the proxy data were interpreted as at or near post-AHP levels.

To further illustrate the temporal variations in the timing of hydrologic changes, we pooled all of the data for low and high humidity conditions across 5° latitude bands to produce frequency histograms at a 500 yr timestep, following an approach similar to 29 (main text Fig 3, Fig S7a). These histograms serve to summarize the frequency of evidence for wet conditions as a function of time and indicate a time-transgressive shift from wet to dry conditions at the end of the AHP. However, we note that interpretations of these plots should consider the potential complications associated with deflation of the records at the most northernmost sites, which dried out completely. Deflation will preferentially remove the most recent portion of the record with the effect of biasing preservation towards older sediments and could potentially account for some of these apparent time-transgressive features. However, the limited number of continuous records tend to support these reconstructed trends. We also note that a time-transgressive trend in the timing of the end of the AHP also occurs going from East to West (Fig. S7b). While records from West and Central Africa tend to show a later demise of the AHP, sites to the east (>20° E) occur earlier e.g., 4-5000 cal yr BP. Unlike the records from West Africa, however, the east-west differences in timing cannot be attributed to the truncated nature of the records, providing support the trends apparent in the histograms are robust.

1.6 Changes in the monsoon from the TraCE-21 experiments To investigate the climate dynamics driving the observed changes in the West African Monsoon, we compared our results to transient simulations of the last 21,000 years performed using the Community Climate System Model, version 3 (CCM3) maintained at NCAR (i.e, the TraCE-21 experiments; www.cgd.ucar.edu/ccr/TraCE/). For most of the comparisons, we chose to examine variables from the ~3.7 x 3.7 degree grid cell centered at 9.8° N, and 0°E. This cell is located about 3° north of Lake Bosumtwi, however, we found it to be most representative of climate at Lake Bosumtwi, primarily because the thermal equator is situated too far North in the TraCE simulation, causing the

© 2015 Macmillan Publishers Limited. All rights reserved grid cell centered at 5.6° N, 0°E to primarily, and unrealistically, respond to changes in Southern Hemisphere insolation. The TraCE simulation captures the primary pattern of the seasonal migration of the West African rainbelt (Figure S8), however it does not properly simulate the mid- summer dry season on the Guinea Coast. Instead of passing over the Guinea Coast after the May-June rainy season, the monsoon belt expands northward into the Sahel while precipitation continues, and even reaches a maximum during August on the Guinea Coast. This effect is enhanced by our choice of grid cell situated North of the lake. To better understand the climate dynamics underlying the observed and simulated changes in moisture balance at Lake Bosumtwi during the deglaciation, especially with respect to freshwater forcing in the North Atlantic during Heinrich 1 and the Younger Dryas, we examined the behavior of the African and Tropical Easterly Jets during those intervals. Throughout the TraCE-21 experiment, the speed of the upper-level Tropical Easterly Jet, which is associated with the intensity of the West African Monsoon (Nicholson, 2008), is insensitive to freshwater forcing, instead, closely tracking tracks Boreal summer insolation (TEJ; Figures S9). This association is likely due to the role of the Asian Monsoon in modulating TEJ speed. In contrast, the position of the mid-level African Easterly does respond to freshwater forcing in the North Atlantic, shifting southward and intensifying over the Sahel during periods of cool SSTs in the North Atlantic (Figure S10). This behavior has been observed in previous experiments (Mulitza et al., 2008), and the enhanced moisture export is hypothesized to have contributed to decreases in moisture in the region during these events. In contrast to the deglacial period, simulated precipitation at Lake Bosumtwi does not record the variability observed from the proxy evidence during the Holocene. The lack of a mid-summer dry season in the simulation may explain this data-model mismatch. Focusing on seasonal, rather than annual, precipitation, the peaks in precipitation corresponding to the two rainy seasons at Lake Bosumtwi are observed in the simulation (Fig S11). Precipitation during both the first (MJJ) and second (SO) rainy seasons track local insolation during those months, implicating the role of local insolation on precipitation, and supporting our hypothesis that the late Holocene peak in

© 2015 Macmillan Publishers Limited. All rights reserved precipitation at Lake Bosumtwi, and elsewhere in low-latitude West Africa, was driven in part by an increase in precipitation during the Boreal fall, rather than the summer.

Supplementary References

1 Shanahan, T. M. et al. Paleoclimatic variations in West Africa from a record of late Pleistocene and Holocene lake level stands of Lake Bosumtwi, Ghana. Paleogeogr. Paleoclimatol. Paleoecol. 242, 287-302 (2006). 2 Turner, B. F., Gardner, L. R. & Sharp, W. E. The hydrology of Lake Bosumtwi, a climate-sensitive lake in Ghana, West Africa. Journal of Hydrology 183, 243-261 (1996). 3 Shanahan, T. M., Overpeck, J. T., Sharp, W. E., Scholz, C. A. & Arko, J. A. Simulating the response of a closed-basin lake to recent climate changes in tropical West Africa (Lake Bosumtwi, Ghana). Hydrological Processes 21, 1678- 1691 (2007). 4 Shanahan, T. M. et al. The formation of biogeochemical laminations in Lake Bosumtwi, Ghana, and their usefulness as indicators of past environmental changes. Journal of Paleolimnology 40, 339-355 (2008). 5 Shanahan, T. M. et al. Spatial and temporal variability in sedimentological and geochemical properties of sediments from an anoxic crater lake in West Africa: Implications for paleoenvironmental reconstructions. Paleogeogr. Paleoclimatol. Paleoecol. 374, 96-109, (2013). 6 Shanahan, T. M. et al. Age models for long lacustrine sediment records using multiple dating approaches - An example from Lake Bosumtwi, Ghana. Quaternary Geochronology 15, 47-60, doi:10.1016/j.quageo.2012.12.001 (2013). 7 Shanahan, T. M. et al. Late Quaternary sedimentological and climate changes at Lake Bosumtwi Ghana: New constraints from laminae analysis and radiocarbon age modeling. Palaeogeography, Palaeoclimatology, Palaeoecology (2012). 8 Blaauw, M. & Christen, J. Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Analysis 6, 457-474 (2011). 9 Reimer, P. J. et al. INTCAL09 and MARINE09 radiocabon age calibration curves, 0-50,000 Years cal BP. Radiocarbon 51, 1111-1150 (2009). 10 Christen, J. & Fox, C. A general purpose sampling algorithm for continuous distributions (the t-walk). Bayesian Analysis 4, 263-282 (2010). 11 Talbot, M. R. & Delibrias, G. A new late-Holocene water-level curve for Lake Bosumtwi, Ghana. Earth and Planetary Science Letters 47, 336-344 (1980).

© 2015 Macmillan Publishers Limited. All rights reserved 12 Brooks, K. et al. Late-Quaternary lowstands of lake Bosumtwi, Ghana: evidence from high-resolution seismic-reflection and sediment-core data. Paleogeogr. Paleoclimatol. Paleoecol. 216, 235-249 (2005). 13 Shanahan, T. M. et al. Atlantic Forcing of Persistent Drought in West Africa. Science 324, 377-380, (2009). 14 Lifton, N. A. et al. Addressing solar modulation and long-term uncertainties in scaling secondary cosmic rays for in situ cosmogenic nuclide applications. Earth and Planetary Science Letters 239, 140-161, (2005). 15 Balco, G., Stone, J. O., Lifton, N. A. & Dunai, T. J. A complete and easily accessible means of calculating surface exposure ages or erosion rates from (10)Be and (26)Al measurements. Quaternary Geochronology 3, 174-195, (2008). 16 Schrag, D. P., Hampt, G. & Murray, D. W. Pore fluid constraints on the temperature and oxygen isotopic composition of the glacial ocean. Science 272, 1930-1932, (1996). 17 Shackleton, N. J. The 100,000-year ice-age cycle identified and found to lag temperature, carbon dioxide, and orbital eccentricity. Science 289, 1897-1902, (2000). 18 Tierney, J. E., Russell, J. M., Damste, J. S. S., Huang, Y. & Verschuren, D. Late Quaternary behavior of the East African monsoon and the importance of the Congo Air Boundary. Quaternary Science Reviews 30, 798-807, (2011). 19 Vimeux, F. et al. A 420,000 year deuterium excess record from East Antarctica: Information on past changes in the origin of precipitation at Vostok. Journal of Geophysical Research-Atmospheres 106, 31863-31873, (2001). 20 Lisiecki, L. E. & Raymo, M. E. A Pliocene-Pleistocene stack of 57 globally distributed benthic delta O-18 records (vol 20, art no PA1003, 2005). Paleoceanography 20, (2005). 21 Sachse, D. et al. Molecular Paleohydrology: Interpreting the Hydrogen-Isotopic Composition of Lipid Biomarkers from Photosynthesizing Organisms. Annu. Rev. Earth Planet. Sci. 40, 221-249 (2012). 22 Collins, J. A. et al. Estimating the hydrogen isotopic composition of past precipitation using leaf-waxes from western Africa. Quaternary Science Reviews 65, 88-101 (2013). 23 Maley, J. Late Quaternary climatic changes in the African rain forest: Forest refugia and the major role of sea surface temperature variations. In Paleoclimatology and Paleometeorology: Modern and Past Patterns of Global Atmospheric Circulation (eds M. Leinen & M. Sarnthein) 585-616 (Kluwer Academic Publishers, 1989). 24 Smith, F. A., Freeman, K. H. . Influence of physiology and climate on δD of leaf wax n- alkanes from C3 and C4 grasses. Geochimica et Cosmochimica Acta 70, 1172-1187 (2006). 25 Polissar, P. J., Freeman, K. H. . Effects of aridity and vegetation on plant-wax δD in modern lake sediments. Geochimica et Cosmochimica Acta 74, 5785-5797. (2010). 26 Stuiver, M., Reimer, P. J., Reimer, R. W. CALIB 5.0. [http://calib.qub.ac.uk/calib/manual/%5D. (2005). 27 Hoelzmann, P. et al. in Past climate variability through Europe and Africa 219-256 (Springer, 2004).

© 2015 Macmillan Publishers Limited. All rights reserved 28 Overpeck, J., Anderson, D., Trumbore, S. & Prell, W. The southwest Indian Monsoon over the last 18 000 years. Climate Dynamics 12, 213-225 (1996). 29 Lezine, A. M., Hely, C., Grenier, C., Braconnot, P. & Krinner, G. Sahara and Sahel vulnerability to climate changes, lessons from Holocene hydrological data. Quaternary Science Reviews 30, 3001-3012, (2011). 30 Huffman, G. J., R.F. Adler, D.T. Bolvin, G. Gu, E.J. Nelk. The TRMM Multi- satellite Precipitation Analysis: Quasi-global, multi-year, combined sensor precipitation estimates at fine scale. Journal of Hydrometeorlogy 8, 38-55 (2007). 31 Schefuß, E., Schouten, S. & Schneider, R. Climatic controls on central African hydrology during the past 20,000 years. Nature 437, 1003-1006 (2005). 32 Niedermeyer, E. M. et al. Orbital- and millennial-scale changes in the hydrologic cycle and vegetation in the western African Sahel: insights from individual plant wax δD and δ13C. Quaternary Science Reviews 29, 2996-3005, doi:10.1016/j.quascirev.2010.06.039 (2010). 33 Talbot, M. R. & Johannessen, T. A high resolution paleoclimatic record for the last 27,000 years in tropical West Africa from the carbon and nitrogen isotopic composition of lacustrine organic matter. Earth and Planetary Science Letters 110, 23-37 (1992). 34 Garcin, Y. et al. Hydrogen isotope ratios of lacustrine sedimentary n-alkanes as proxies of tropical African hydrology: Insights from a calibration transect across Cameroon. Geochimica et Cosmochimica Acta 79, 106-126, (2012). 35 Gasse, F., Téhet, R., Durand, A., Gibert, E., Fontes, J.C. The arid-humid transition in the Sahara and the Sahel during the last deglaciation. Nature 345, 141-146 (1990). 36 Gasse, F. Diatom-inferred salinity and carbonate oxygen isotopes in Holocene waterbodies of the western Sahara and Sahel (Africa). Quaternary Science Reviews 21, 737-767 (2002). 37 Gilbert, A. M., Conrad, G., De Deckker, P., Fontes, J.C., Gasse, F., Kassir, A. Retour des conditions humides au Tardiglaciaire au Sahara septentrional (Sebkha Mellala, Algerie). Bulletin de la Société Géologique de France 6, 497-504 (1990). 38 Fontes, J. C. et al. Freshwater to marine-like environments from Holocene lakes in northern Sahara. Nature 317, 608-610, (1985). 39 Gasse, F., Fontes, J.C., Plaziat, J.C., Carbonnel, P., Kaczmarska, P., De Deckker, P., Soulié-Marsche, I., Callot, Y. Biological remains, geochemistry and stable isotopes for the reconstruction of environmental and hydrological changes in the Holocene lakes from North Sahara. Palaeogeography, Palaeoclimatology, Palaeoecology 60, 1-46 (1987). 40 Kuhlmann, H., Meggers, H., Freudenthal, T. & Wefer, G. The transition of the monsoonal and the N Atlantic climate system off NW Africa during the Holocene. Geophysical Research Letters 31, L22204, (2004). 41 McGee, D., Demenocal, P. B., Winckler, G., Stuut, J. B. W. & Bradtmiller, L. I. The magnitude, timing and abruptness of changes in North African dust deposition over the last 20,000yr. Earth Planet Sc Lett 371-372, 163-176, (2013). 42 deMenocal, P., Ortiz, Joseph, Guilderson, Tom, Sarnthein, Michael. Coherent High- and Low-Latitude Climate Variability During the Holocene Warm Period. Science 288, 2198-2202 (2000).

© 2015 Macmillan Publishers Limited. All rights reserved 43 Multiza. Sahel megadroughts triggered by glacial slowdowns of Atlantic meridional overturning. 1-30 (2008). 44 Lezine, A.-M. & Chateauneuf, J.-J. Peat in the "Niayes" of Senegal: depositional environment and Holocene evolution. Journal of African Earth Sciences (and the Middle East) 12, 171-179 (1991). 45 Deynoux, M. et al. Stratigraphie séquentielle en milieu désertique. Exemple de l'erg Akchar en Mauritanie occidentale. Comptes Rendus Académie des Sciences, Paris II 317, 1199-1205 (1993). 46 Kocurek, G., Havholm, K. G., Deynoux, M. & Blakey, R. C. Amalgamated accumulations resulting from climatic and eustatic changes, Akchar Erg, Mauritania. Sedimentology 38, 751-772 (1991). 47 Lézine, A.-M. Paléoenvironnements végétaux d'Afrique nord tropicale depuis 12000 BP, University Aix-Marseille, (1987). 48 Lézine, A.-M. & Casanova, J. Pollen and hydrological evidence for the interpretation of past climates in tropical West Africa during the Holocene. Quaternary Science Reviews 8, 45-55 (1989). 49 Lézine, A.-M., Casanova, J. & Hillaire-Marcel, C. Across an early Holocene humid phase in western Sahara: Pollen and isotope stratigraphy. Geology 18, 264- 267 (1990). 50 Delibrias, G., Petit-Maire, N. & Fabre, J. Ages des dépôts lacustres récents de la région de Taoudenni-Trhaza (Sahara malien). Comptes Rendus de l'Académie des Sciences, Paris 299, 1343-1346 (1984). 51 Fabre, J. & Petit-Maire, N. Holocene climatic evolution at 22–23 N from two palaeolakes in the Taoudenni area (northern Mali). Palaeogeography, Palaeoclimatology, Palaeoecology 65, 133-148 (1988). 52 Petit-Maire, N. Taoudenni basin (Mali), Holocene palaeolimnology and environments. Würzburger Geographische Arbeiten 69, 45-52 (1988). 53 Hillaire-Marcel, C., Riser, J., Rognon, P., Petit-Maire, N., Rosso, J.C., Soulie- Marche, I. Radiocarbon chronology of Holocene hydrologic changes in northeastern Mali. Quaternary Research 20, 145-164 (1983). 54 Petit-Maire, N. Palaeoclimates in the Sahara of Mali. A multidisciplinary Study. Episodes 9, 1-18 (1986). 55 Elenga, H., Maley, J., Vincens, A. & Farrera, I. in Past climate variability through Europe and Africa 181-198 (Springer, 2004). 56 Lespez, L. et al. High-resolution fluvial records of Holocene environmental changes in the Sahel: the YamE River at Ounjougou (Mali, West Africa). Quaternary Science Reviews 30, 737-756, (2011). 57 Ballouche, A. & Neumann, K. A new contribution to the Holocene vegetation history of the West African Sahel: pollen from Oursi, Burkina Faso and charcoal from three sites in northeast Nigeria. Vegetation History and Archaeobotany 4, 31-39 (1995). 58 Cremaschi, M., di Lernia, S. Holocene climatic changes and cultural dynamics in the Libyan Sahara. African Archaeological Review 16, 211-243 (1999). 59 Cremaschi, M., Zerboni, A. Early to middle holocene landscape exploitation in a drying environment: two case studies compared from the central Sahara (SW Fezzan, Libya). Comptes Rendus Geoscience 341, 689-702 (2009).

© 2015 Macmillan Publishers Limited. All rights reserved 60 Baumhauer, R. Palaeolakes of the south central Sahara – problems of palaeoclimatological interpretation. Hydrobiologia 214, 347-357 (1991). 61 Servant, M. Sequences continentales et variations climatiques: evolution du bassin du Tchad au Cenozoique superieur. 573 (Orstom, 1983). 62 Grunert, J., Baumhauser, R., Vökel, J. Lacustrine sediments and Holocene climates in the southern Sahara: the example of paleolakes in the Grand Erg of Bilma (Zoo Baba and Dibelle, eastern Niger). Journal of African Earth Sciences 12, 133-146 (1991). 63 Garba, Z., Durand, A., Lang, J. Enregistrement sédimentaire des variations de la dynamique éolienne pendant la transition Tardiglaciaire/Holocène à la limite Sahara/Sahel (Termit, bassin du lac Tchad). Zeitschrift für Geomorphologie N.F. Suppl. Bd 103, 159-178 (1996). 64 Waller, M. P., Street-Perrott, F. A. & Wang, H. Holocene vegetation history of the Sahel: pollen, sedimentological and geochemical data from Jikariya Lake, north-eastern Nigeria. Journal of Biogeography 34, 1575-1590, (2007). 65 Salzmann, U., Waller, M. The Holocene vegetational history of the Nigerian Sahel based on multiple pollen profiles. Review of Palaeobotany and Palynology 100, 39-72 (1998). 66 Holmes, J. A., Street-Perrott, F.A., Allen, M.J., Fothergill, P.A., Harkness, D.D., Kroon, D., Perrott, R.A. Holocene palaeolimnology of Kajemarum Oasis, Northern Nigeria: an isotopic study of ostracods, bulk carbonate and organic carbon. Journal of the Geological Society, London 154, 311-319 (1997). 67 Street-Perrott, F. A., Holmes, J.A., Waller, M.P., Allen, M.J., Barber, N.G.H., Fothergill, P.A., Harkness, D.D., Ivanovich, M., Kroon, D., Perrott, R.A. Drought and dust deposition in the West African Sahel: a 5500-year record from Kajemarum Oasis, northeastern Nigeria. The Holocene 10, 293-302 (2000). 68 Gumnior, M., Preusser, F. Late Quaternary river development in the southwest Chad Basin: OSL dating of sediment from the Komadugu palaeofloodplain (northeast Nigeria). Journal of Quaternary Science 22, 709–719 (2007). 69 Hoelzmann, P., Kruse, H.J., Rottinger, F. Precipitation estimates for the eastern Saharan palaeomonsoon based on a water balance model of the West Nubian Palaeolake Basin. Global and Planetary Change 26, 105-120 (2000). 70 Hoelzmann, P., Keding, B., Berke, H., Kröpelin, S., Kruse, H.J. Environmental change and archaeology: lake evolution and human occupation in the Eastern Sahara during the Holocene. Palaeogeography, Palaeoclimatology, Palaeocology 169, 193-217 (2001). 71 Pachur, H. J., Kröpelin, S., Hoelzmann, P., Goschin, M., Altmann, N. Late quaternary fluvio-lacustrine environments of Western Nubian. Berliner Geowissenschaftliche Abhandlungen (A) 120, 203-260 (1990). 72 Pachur, H. J., Hoelzmann, P. Palaeoclimatic implications of Late Quaternary Lacustrine sediments in Western Nubia, Sudan. Quaternary Research 36, 257-276 (1991). 73 Gasse, F. o. & Van Campo, E. Abrupt post-glacial climate events in West Asia and North Africa monsoon domains. Earth and Planetary Science Letters 126, 435-456 (1994).

© 2015 Macmillan Publishers Limited. All rights reserved 74 Servant, M. & Servant-Vildary, S. L'environnement quaternaire du bassin du Tchad. The Sahara and the Nile, 133-162 (1980). 75 Eggermont, H. et al. Aquatic community response in a groundwater-fed desert lake to Holocene desiccation of the Sahara. Quaternary Science Reviews 27, 2411-2425, (2008). 76 Kroepelin, S. et al. Climate-driven ecosystem succession in the Sahara: The past 6000 years. Science 320, (2008). 77 Donner, J. The geology of a playa in Farafra, Western Desert of Egypt. Palaeoecology of Africa 25, 121-131 (1998). 78 Donner, J. J., Ashour, M.M., Embabi, N.S., Siiriäinen, A. The Quaternary geology of a playa in Farafra, Western Desert of Egypt. Annales Academiae Scientiarum Fennicae. Geologica-Geographica 160, 49-112 (1999). 79 Hassan, F. A., Barich, B., Mahmoud, M., Hemdan, M.A. Holocene playa deposits of Farafra Oasis, Egypt, and their palaeoclimatic and geoarchaeological significance. Geoarchaeology: an international journal 16, 29-46 (2001). 80 Brookes, I. A. Early Holocene basinal sediments of the Dakhleh oasis region, South central Egypt. Quaternary Research 32, 139-152 (1989). 81 Brookes, I. A. Geomorphology and Quaternary geology of the Dakhla Oasis Region, Egypt. Quaternary Science Reviews 12, 529-552 (1993). 82 Haynes, C. V. in Prehistory of Arid North Africa (ed A.E. Close) 69-85 (Southern Methodist University Press, 1987). 83 Haynes, C. V. Geochronology and climat change of the Pleistocene-Holocene transition in the Darb el Arba’in Desert, Eastern Sahara. Geoarchaeology 16, 119- 141 (2001). 84 Pachur, H. J., Wünnemann, B. Reconstruction of the palaeoclimate along 30° E in the eastern Sahara during the Pleistocene/ Holocene transition. Palaeoecology of Africa 26, 1-32 (1996). 85 Ritchie, J. C., Haynes, C.V. Holocene vegetation zonation in eastern Sahara. Nature 330, 645-647 (1987). 86 Szabo, B. J., Haynes, C.V., Maxwell, T.A. Ages of Quaternary pluvial episodes determined by uranium-serie and radiocarbon dating of lacustrine deposits of Eastern Sahara. Palaeogeography, Palaeclimatology, Palaeoecology 113, 227- 242 (1995). 87 Abell, P. I., Hoelzmann, P., Pachur, H.J. Stable isotopes ratios of gastropod shells and carbonate sediments of NW Sudan ans palaeoclimatic indicators. Palaeoecology of Africa 26, 33-52 (1996). 88 Abell, P. I., Hoelzmann, P. Holocene paleoclimates in northwestern Sudan: stable isotope studies on molluscs. Global and Planetary Change 26, 1-12 (2000). 89 Ritchie, J. C., Eyles, C.H., Haynes, C.V. Sediment and pollen evidence for an early to mid-Holocene humid period in the eastern Sahara. Nature 314, 352-355 (1985). 90 Hoelzmann, P., Schwab, A., Robers, N., Cooper, P., Burgess, A. Hydrological response of an east-Saharan palaeolake (NW Sudan) to early-Holocene climate. The Holocene 20, 537-549 (2010). 91 Pachur, H. J., Hoelzmann, P. Late Quaternary palaeoecology and palaeoclimates of the eastern Sahara. Journal of African Earth Sciences 30, 929-939 (2000).

© 2015 Macmillan Publishers Limited. All rights reserved 92 Bubenzer, O., Besler, H., Hilgers, A. Filling the gap: OSL data expanding 14C chronologies of Late Quaternary environmental change in the Libyan Desert. Quaternary International 175, 41–52 (2007). 93 Pachur, H. J., Röper, H.P. The libyan (Western) desert and northern Sudan during the late Pleistocene and holocene Berliner Geowissenschaftliche Abhandlungen (A) 50, 249-284 (1984). 94 Ayliffe, D., Williams, M.A.J., Sheldon, F. Stable isotope and ocygen isotopic composition of early-Holocene gastropods from Wadi Mansurab, north-central Sudan. The Holocene 6, 157-169 (1996). 95 Williams, M. A. J., Adamson, D.A. in The Sahara and the Nile (ed M.A.J. Williams, Faure, H.) 281-304 (Balkema, 1981). 96 Cremaschi, M., Zerboni, A., Spötl, C., Felleti, F. The calcareous tufa in the Tadrart Acacus Mt. (SW Fezzan, Lybia). An early holocene palaeoclimate archive in the central sahara. Palaeogeography, Palaeoclimatology, Palaeoecology 287, 81-94 (2010). 97 Barker, P., Telford, R., Gasse, F. & Thevenon, F. Late Pleistocene and Holocene palaeohydrology of Lake Rukwa, Tanzania, inferred from diatom analysis. Palaeogeography, Palaeoclimatology, Palaeoecology 187, 295-305 (2002). 98 Barker, P., Williamson, D., Gasse, F. & Gibert, E. Climatic and volcanic forcing revealed in a 50,000-year diatom record from Lake Massoko, Tanzania. Quaternary Research 60, 368-376 (2003). 99 Costa, K., Russell, J., Konecky, B. & Lamb, H. Isotopic reconstruction of the African Humid Period and Congo Air Boundary migration at Lake Tana, Ethiopia. Quaternary Science Reviews 83, 58-67, doi:10.1016/j.quascirev.2013.10.031 (2014). 100 Gasse, F. Hydrological changes in the African tropics since the Last Glacial Maximum. Quaternary Science Reviews 19, 189-211 (2000). 101 Gasse, F., Rognon, P. & Street, F. Quaternary history of the Afar and Ethiopian Rift lakes. The Sahara and the Nile (MAJ Williams and H. Faure, Editors). . 361- 400 (1980). 102 Chalié, F. & Gasse, F. Late Glacial-Holocene diatom record of water chemistry and lake level change from the tropical East African Rift Lake Abiyata (Ethiopia). Palaeogeography, Palaeoclimatology, Palaeoecology 187, 259-283 (2002). 103 Tierney, J. E. & deMenocal, P. B. Abrupt shifts in Horn of Africa hydroclimate since the Last Glacial Maximum. Science Express published online October 10 2013 (2013). 104 Jung, S., Davies, G., Ganssen, G. & Kroon, D. Stepwise Holocene aridification in NE Africa deduced from dust-borne radiogenic isotope records. Earth and Planetary Science Letters 221, 27-37, (2004). 105 Benvenuti, M. et al. The Ziway-Shala lake basin (Main Ethiopian rift, Ethiopia): a revision of basin evolution with special reference to the Late Quaternary. Journal of African Earth Sciences 35, 247-269 (2002). 106 Gasse, E. & Street, F. Late Quaternary lake-level fluctuations and environments of the northern Rift Valley and Afar region (Ethiopia and Djibouti). Palaeogeography, Palaeoclimatology, Palaeoecology 24, 279-325 (1978).

© 2015 Macmillan Publishers Limited. All rights reserved 107 Gillespie, R., Street-Perrott, F. A. & Switsur, R. Post-glacial arid episodes in Ethiopia have implications for climate prediction. Nature 306, 680-683 (1983). 108 Lamb, A., Leng, M., Lamb, H. & Mohammed, M. A 9000-year oxygen and carbon isotope record of hydrological change in a small Ethiopian crater lake. Holocene 10, 167-177 (2000). 109 Jolly, D., Bonnefille, R. & Roux, M. Numerical interpretation of a high resolution Holocene pollen record from Burundi. Palaeogeography, Palaeoclimatology, Palaeoecology 109, 357-370 (1994). 110 Aucour, A., Bonnefille, R. & Hillaire-Marcel, C. Sources and accumulation rates of organic carbon in an equatorial peat bog (Burundi, East Africa) during the Holocene: carbon isotope constraints. Paleogeogr. Paleoclimatol. Paleoecol. 150, 179-189 (1999). 111 Bonnefille, R. et al. Glacial/interglacial record from intertropical Africa, high resolution pollen and carbon data at Rusaka, Burundi. Quaternary Science Reviews 14, 917-936 (1995). 112 Taylor, D. Late Quaternary pollen records from two Ugandan mires: Evidence for environmental change in the Rukiga Highlands of southwest Uganda. Palaeogeography, Palaeoclimatology, Palaeoecology 80, 283-300 (1990). 113 Beuning, K. & Russell, J. Vegetation and sedimentation in the Lake Edward Basin, Uganda-Congo during the late Pleistocene and early Holocene. Journal of Paleolimnology 32, 1-18 (2004). 114 Owen, R., Barthelme, J., Renaut, R. & Vincens, A. Palaeolimnology and archaeology of Holocene deposits north-east of Lake Turkana, Kenya. (1982). 115 Beuning, K., Kelts, K., Russell, J. & Wolfe, B. Reassessment of Lake Victoria - Upper Nile River paleohydrology from oxygen isotope records of lake-sediment cellulose. Geology 30, 559-562 (2002). 116 Stager, J., Mayewski, P. & Meeker, L. Cooling cycles, Heinrich event 1, and the desiccation of Lake Victoria. Paleogeogr. Paleoclimatol. Paleoecol. 183, 169-178 (2002). 117 Street-Perrott, F. & Perrott, R. Abrupt climate fluctuations in the tropics: the influence of Atlantic Ocean circulation. Nature 343, 607-612 (1990). 118 Tierney, J. et al. Northern Hemisphere controls on tropical southeast African climate during the past 60,000 years. Science 322, 252 (2008). 119 Stager, J. Environmental changes at Lake Cheshi, Zambia since 40,000 years BP. Quaternary research (USA) (1988). 120 Salzmann, U. Are modern savannas degraded forests ? - A Holocene pollen record from the Sudanian vegetation zone of NE Nigeria. Vegetation history and archaeobotany 9, 1-15 (2000). 121 Salzmann, U., Hoelzmann, P. & Morczinek, I. Late Quaternary climate and vegetation of the Sudanian zone of northeast Nigeria. Quaternary Research 58, 73-83 (2002). 122 Baumhauer, R., Schulz, E. & Pomel, S. in Paleoecology of Quaternary Drylands 31-45 (Springer, 2004). 123 Salzmann, U. & Hoelzmann, P. The Dahomey Gap: an abrupt climatically induced rain forest fragmentation in West Africa during the late Holocene. Holocene 15, 190-199, (2005).

© 2015 Macmillan Publishers Limited. All rights reserved 124 Tossou, M. G., Akoegninou, A., Ballouche, A., Sowunmi, M. A. & Akpagana, K. The history of the mangrove vegetation in Benin during the Holocene: A palynological study. Journal of African Earth Sciences 52, 167-174, (2008). 125 Sowunmi, M. A. in Past climate variability through Europe and Africa 199- 218 (Springer, 2004). 126 Fredoux, A. Evolution de la vegetation dans le delta de l'Agneby (Cote d'Ivoire). Travaux et documents de geographie tropicale, 117-134 (1983). 127 Giresse, P., Maley, J. & Brenac, P. Late Quaternary palaeoenvironments in the Lake Barombi Mbo (West Cameroon) deduced from pollen and carbon isotopes of organic matter. Paleogeogr. Paleoclimatol. Paleoecol. 107, 65-78 (1994). 128 Lebamba, J., Vincens, A. & Maley, J. Pollen, vegetation change and climate at Lake Barombi Mbo (Cameroon) during the last ca. 33 000 cal yr BP: a numerical approach. Climate of the Past 8, 59-78, (2012). 129 Maley, J. The African rain forest vegetation and paleoenvironments during the late Quaternary. Climatic Change 19, 79-98 (1991). 130 Weldeab, S., Lea, D. W., Schneider, R. R. & Andersen, N. 155,000 Years of West African Monsoon and Ocean Thermal Evolution. Science 316, 1303-1307, (2007). 131 Lézine, A. Timing of vegetation changes at the end of the Holocene Humid Period in desert areas at the northern edge of the Atlantic and Indian monsoon systems. Comptes rendus-Géoscience 341, 750-759 (2009). 132 Vincens, A. et al. Forest response to climate changes in Atlantic Equatorial Africa during the last 4000 years BP and inheritance on the modern landscapes. Journal of Biogeography 26, 879-885 (1999). 133 Zogning, A., Giresse, P., Maley, J. & Gadel, F. o. The Late Holocene palaeoenvironment in the Lake Njupi area, west Cameroon: implications regarding the history of Lake Nyos. Journal of African Earth Sciences 24, 285- 300 (1997). 134 Kadomura, H. in Savannization Processes in Tropical Africa II (ed H. Kadomura, Kiyonaga, J.) 47–85 (Tokyo Metropolitan University, 1994). 135 Moeyersons, J. Geomorphological processes and their paleoenvironmental significance at the Shum Laka cave (Bamenda, western Cameroon). Palaeogeography, Palaeoclimatology, Palaeoecology 133, 103-116 (1997). 136 Tamura, T. Late Quaternary landscape evolution in the west Cameroon Highlands and the Adamaoua Plateau. Paysages Quaternaires de L‚ÄôAfrique Centrale Atlantique. Paris: Editions de l‚ÄôOrstom, 298-313 (1990). 137 Nguetsop, V. et al. Past environmental and climatic changes during the last 7200 cal yr BP in Adamawa plateau (Northern-Cameroun) based on fossil diatoms and sedimentary carbon isotopic records from Lake Mbalang. Climate of the Past 7, 1371-1393 (2011). 138 Vincens, A., Buchet, G. & Servant, M. Vegetation response to the "African Humid Period" termination in central Cameroon (7 N)- new polen insight from Lake Mbalang. Climate of the Past 6, 281-294 (2010). 139 Giresse, P., Maley, J. & Kossoni, A. Sedimentary environmental changes and millennial climatic variability in a tropical shallow lake (Lake Ossa, Cameroon) during the Holocene. Paleogeogr. Paleoclimatol. Paleoecol. 218, 257-285 (2005).

© 2015 Macmillan Publishers Limited. All rights reserved 140 Reynaud-Farrera, I., Maley, J. & Wirrmann, D. Végétation et climat dans les forêts du Sud-Ouest Cameroun depuis 4770 ans BP: analyse pollinique des sédiments du Lac Ossa. Comptes Acad Sci Paris 322, 749-755 (1996). 141 Ngomanda, A., K. Neumann, A. Schweizer, J. Maley. Seasonality change and the third millennium BP rainforest crisis in Central Africa: a high resolution pollen profile from Nyabessan, southern Cameroon. Quaternary Research 71, 307-318 (2009). 142 Brncic, T. M., Willis, K. J., Harris, D. J. & Washington, R. Culture or climate? The relative influences of past processes on the composition of the lowland Congo rainforest. Philos T R Soc B 362, 229-242, (2007). 143 Brncic, T. M., Willis, K. J., Harris, D. J., Telfer, M. W. & Bailey, R. M. Fire and climate change impacts on lowland forest composition in northern Congo during the last 2580 years from palaeoecological analyses of a seasonally flooded swamp. The Holocene 19, 79-89 (2009). 144 Laraque, A. et al. Reconnaissance scientifique du lac Télé (Nord-Congo) — Premiers résultats et interprétations. Comptes Rendus de l'Académie des Sciences - Series IIA - Earth and Planetary Science 325, 49-56 (1997). 145 Ngomanda, A. et al. Vegetation changes during the past 1300 years in western equatorial Africa: a high-resolution pollen record from Lake Kamalete, Lope Reserve, Central Gabon. Holocene 15, 1021-1031, (2005). 146 Elenga, H. Végétation et climat du Congo depuis 24 000 ans B.P. : analyse palynologique de séquences sédimentaires du Pays Bateke et du littoral, Université de Droit d'Economie et des Sciences d'Aix-Marseille 3, (1992). 147 Elenga, H. et al. Diagramme pollinique holocène du lac Kitina (Congo): mise en évidence de changements paléobotaniques et paléoclimatiques dans le massif forestier du Mayombe. Compte Rendu de l’Académie des Sciences 323, 403-410 (1996). 148 Vincens, A., Schwartz, D., Bertaux, J., Elenga, H. & de Namur, C. Late Holocene climatic changes in western equatorial Africa inferred from pollen from Lake Sinnda, southern Congo. Quaternary Research 50, 34-45 (1998). 149 Elenga, H. et al. Le marais estuarien de la Songolo (Sud Congo) a l'Holocene moyen et recent. Bulletin de la Société Géologique de France 172, 359-366 (2001). 150 Elenga, H., Schwartz, D. & Vincens, A. Pollen evidence of late Quaternary vegetation and inferred climate changes in Congo. Palaeogeography, Palaeoclimatology, Palaeoecology 109, 345-356 (1994).

6°32'N 6 1 2 4 3 5 70 m

50 m 30 m

6°27'N

90 0 1 2 km

Fig. 1. Map of the Lake Bosumtwi crater. Bathymetry (solid Figuregrey S1.lines, Bathymetric 10 m contour) map from of theBrooks Lake et Bosumtwi al., [2005]. impact Dashed crater. Circlesline indicatesindicate thecrater locations rim. Inset of showscoring thesites location used in of this Lake study Bosumtwi1 in southern Ghana, West Africa.

North African Dust ODP658C

Senegal River GeoB9508-5 Lake Bosumtwi

4.1

Congo Fan 3.2 GeoB6518 2.3

1.4 mm/day

0.5

18 b. Longitude: 1.25° W 14

10 Bosumtwi Latitude N Latitude 6

2 Eq Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Figure S2. Study locations and precipitation over North Africa. a. Mean annual precipitation (mm/day) for the period between 1998-2010 from the Tropical Rainfall Measur- ing Mission (TRMM) v. 6 30. Diamond, site showing an abrupt mid-Holocene termination of the AHP: ODP658c(20°45’N, 18°35’W).42. Squares, sites showing a two-stage collapse of the AHP: Red square, study area, Lake Bosumtwi (6°30’N, 1°25’W), Congo River delta core GeoB6518-1 (5°35.5’S, 11°13.3’E) 31, Senegal River core GeoB9508-S (15°29.9’N, 17°56.88’W)32. b. Hovmöller diagram of the seasonal change in precipitation as a function of latitude at the longitude of Lake Bosumtwi (1°25’W) for 2003 from the Tropical Rainfall Measuring Mission (TRMM) v. 6 dataset.

© 2015 Macmillan Publishers Limited. All rights reserved Figure S3. Age depth model for the composite Lake Bosumtwi sediment record (in RMCD standardized depth). Symbols indicate calibrated age distributions for individual radiocarbon ages. The solid black line indicates the optimal age model determined using the bayesian age-depth modeling software BACON8. Redrawn from 7.

c. -60 m −50

−40 (‰) −30 corr.

δ D −20

−10 d.

Lake level interpretations

water balance > today water balance < today

lake level uncertain sedimentological evidence of deep conditions

Radiocarbon ages in-situ surface exposure age paleoterrace laminated deep water sediments deep water silts highly oxidized laminated sediments

Deep water core stratigraphy unlaminated, ne grained, organic rich silts

nely laminated sediments Figure S4. Comparison between leaf wax isotopes and paleolake level reconstruction for Lake Bosumtwi. a. Summer (June-August) insolation changes at 6.5°N. b. Changes in lake level reconstructed from radiocarbon dated terraces and lacustrine silts preserved on the crater walls and identified below the modern lake level. Data was synthesized from previous workers 1,11,12 and revised using new in-situ 14C production rates14. c. Hydrogen isotope composition of C31 n-alkanes adjusted for changes in vegetation (grey) and vegetation and ice volume (black) (SOM 1.4.3) indicating changes in precipitation. Shading reflects 68% (dark) and 95% (light) uncertainties in the reconstruction, based on analytical and age model errors. d. Idealized core stratigraphy showing the time interval of unit S1, which is organic-rich, unlaminated and dominated by the remains of nitrogen fixing blue green algae indicative of eutrophic conditions33.

δ D −155 −150 −145

−140 −34 −135 −32

−30 b. −28 (‰) −26 wax C

−24 13 δ −22

−20

−50 −18

−45 −16 −40 c. −35 (‰) −30 −25 −20

wax-corrected −15 δ D Veg. corrected −10 −5 Veg. + Ice volume 0 corrected

20 18 16 14 12 10 8 6 4 2 0 Age (ka BP) Figure S5. Correction of the δDwax precipitation record for changes in ice volume and vegetation. a. Measured δDwax values for the C31 n-alkanes in 13 Lake Bosumtwi. b. Measured δ Cwax for the C31 n-alkanes. c. A comparison between the δDwax record after correction for the influence of vegetation type on apparent fractionation between leaf wax and source water (grey) and after a correction for changes in the isotopic composition of source water due to ice volume (blue).

a. 1 2

3 36 4 19 37 5 10 121314 20 11 39 38 6 21 22 33 35 15 25 40 20˚ 7 23 41 43 20˚ 26 24 42 44 8 9 18 46 17 27 29 34 45 49 28 30 50 52 31 48 3267 53 66 47 51 69-75 79 82 54 80 68 81 56 55 16 77 83 85 86 84 57 60 78 87 88 90 58 62 0˚ 89 76 59 61 93 94 63 0˚ 92 96 91 95 64 65

-20˚ 0˚ 20˚ 40˚

-20˚ 0˚ 20˚ 40˚ b.

20˚ 20˚

0 1 2 3 4 15 5

10 cal. kyr BP 0˚ Proxy data 0˚

pollen hydro δDwax dust arid intermediate humid no data

-20˚ 0˚ 20˚ 40˚ Figure S6. a. Sites used in the reconstruction of past hydrological changes across North Africa. See Table S1 for reference number locations. b. Clock wheel diagrams synthesizing the evolution of reconstructed changes in hydrology over the last 20,000 years from the published literature. Colors indicate paleohydrologic or paleovegetation conditions; white indicates no data at a given site. Sites with grey symbols are designated here as East Africa and are excluded from the histograms in Fig. S7a. See SOM sec. 1.5 and SOM Table S1 for details and references on individual records.

Latitude bin % Inferred status: wet moderate Longitude bin % Inferred status: wet moderate

100 100 75 75 >10W >25˚ N 50 >10˚ W 50 25 25 0 0 100 100 75 75 20-25˚ N 50 0-10˚ W 50 25 25 0 0 100 100 75 75 0-10E 15-20˚ N 50 0-10˚ E 50 25 25 0 0 100 100 75 75 10-15E 10-15˚ N 50 10-15˚ E 50 25 25 0 0 100 100 75 75 15-20E 5-10˚ N 50 15-20˚ E 50 25 25 0 0 100 100 0-5˚ N 75 75 20-30E 20-30˚ E 50 50 25 25 0 0 100 100 75 75 >30E <0˚ S 50 >30˚ E 50 25 25 0 0 02468101214161820 02468101214161820

Age (x1000 cal yr BP) Age (x1000 cal yr BP)

Figure S7. Histograms of wet conditions during the AHP across North Africa. a. Same as in main text Figure 3 but with the sites from East Africa excluded. b. Histograms in longitudinal bins going from west to east across the study area. Dark blue: maximum wet conditions. Light blue: moder- ately wet conditions.

20oN 8

10oN 6

0o 4 mm/day

10oS 2

20oS 0 15oW 0 o 15oE 30 oE 45 oE

20o N b.

10o N

o 0 Feb Apr Jun Aug Oct Dec Month

Figure S8. Model simulation of precipitation seasonality over West Africa. a.Simulated mean daily precipitation over tropical and subtropical Africa during the most recent 1000 years of the TraCE-21 simulations (991 – 1990 AD). b. Latitudinal progression of summer monsoon precipitation in West Africa as a function of month, at 0°E. The latitude of Lake Bosumtwi (6.5°N) is shown by the black dashed line; the mid-point latitude (9.8°N) of the grid cell used for comparison in this study shown by the red dashed line.

30oN −1 30oN −1 20 20

o o 15 N 10 15 N 10

0o 0 0o 0

15oS −10 15oS −10

−20 −20 30oS 30oS etrywn pes(m speeds wind s Westerly (m speeds wind s Westerly −30 −30 o W o o 36o 54o o W o o 36o 54o 18 0 18 E E E 18 0 18 E E E

c. Zonal winds 100−200 mb(YD) d. Zonal winds 100−200 mb(H1−BA) 30 10 ) )

30oN −1 30oN −1 20 5 o o 15 N 10 15 N

0o 0 0o 0

15oS −10 15oS −5 −20 30oS 30oS etrywn pes(m speeds wind s Westerly (m speeds wind s Westerly −30 −10 o W o o 36o 54o o W o o 36o 54o 18 0 18 E E E 18 0 18 E E E ) -1 10 450

e. ) -2

440 5

430 0

420 JJA insolation (W m JJA

Easterly Wind Speed (m s Easterly Wind -5 20 16 12 8 4 0

Age (ka)

Figure S9. Speed and velocity of the Tropical Easterly Jet for June-August (JJA) during the deglaciation. A-D) Simulated zonal wind speeds, averaged over 100-200 mb, for A) Heinrich 1 (17 to 14.7 ka), B) Bolling-Allerod (14.6 to 13 ka) C) Younger Dryas (12.9 to 11.5 ka) and D) The difference between A and B. E) Simulated average JJA zonal wind speed over the past 20 kyr, averaged over 0 to 15°N, -20 to 20°E, and 100 to 200 mb shown in black, with average JJA insolation at 6.5 N in red.

100 20 200 ) −1 300 10 400

500 0

600 rsue(mb) Pressure −10 700 etrywn pes(m speeds wind s Westerly 800 −20 900

−30 −10 0 10 20 30 40 −10 0 10 20 30 40

Zonal winds along Greenwhich (YD) Zonal wind anomalies along Greenwhich (H1−BA) 10

100 8 )

200 6 −1

300 4

400 2

500 0

600 −2 rsue(mb) Pressure

700 −4

800 −6 etrywn pe nmle (m s anomalies speed wind Westerly 900 −8

−10 −10 0 10 20 30 40 −10 0 10 20 30 40 Latitude Latitude

Figure S10. Simulated zonal wind speeds between 10°S and 40°N along the Greenwhich Meridian, for Heinrich 1 (H1; 17 to 14.7 ka), the Bolling-Allerod (BA; 14.6 to 13 ka), the Younger Dryas (YD; 12.9 to 11.5 ka) and the difference between H1 and BA.

26 430 ) -2

22 rcp(cm/month) Precip noain(Wm Insolation 410

14 390 20 16 12 8 4 0

Age(ka)

Figure S11. Differences in monthly precipitation variability driven by changing insolation over the Holocene. Simulated seasonal precipitation and insolation at the Lake Bosumtwi over the past 22,000 years. Mean monthly precipitation for May, June and July are shown in blue, and mean monthly precipitation for September and October are shown in green. Both precipitation curves are smoothed with an 11-year running mean. Seasonal mean insolation at 6.5°N for the corresponding months are shown in cyan and bright green.

The time-transgressive termination of the African Humid

Period

Timothy M. Shanahan, Nicholas P. McKay, Konrad A. Hughen, Jonathan T. Overpeck,

Bette Otto-Bliesner, Clifford W. Heil, John King, Christopher A. Scholz, John Peck

Supplementary Table

Table S1. Locations of paleohydrological sites across North Africa.

number site location lat lon proxy reference 1 Sebkha mellala Algeria 32.18 5.20 1,2 27,35-37 2 Hassi el Mejnah Algeria 31.67 2.50 1,2,3 27,36,38,39 - 40 3 GeoB5546 Atlantic 27.54 13.74 4 - 41 4 OC437-7 GC37 Atlantic 26.82 15.12 4 - 41 5 OC437-7 GC49 Atlantic 23.21 17.85 4 - 42 6 ODP 658C Cape Blanc 20.75 18.58 4 - 41 7 OC437-7 GC68 Atlantic 19.36 17.28 4 - 32,43 8 GeoB9508-5 Senegal 15.50 17.95 5,7 - 44 9 Diogo Senegal 15.26 16.80 6 - Erg Akchar Mauritania 21.98 27,45,46 10 13.32 2 - Chemchane Mauritania 20.93 27,47-49 11 12.22 7 12 Agorgott Mali 22.65 -4.02 2 27,50-52 13 Wadi Haijad Mali 22.57 -3.67 1,2,3 27,36,53,54 14 Tagnout-Chaggaret Mali 21.22 -0.68 2 27,53 15 Ine Kousamene Mali 20.60 -0.85 2 27,53 126 16 Agneby Ivory Coast 5.33 -4.25 6 56 17 Ounjougou Mali 14.60 -3.07 2 Burkina 57

Oursi 14.65 -0.13 18 Faso 6 19 erg Uan Kasa Libya 25.76 10.89 2,3 58,59 20 Tadrart Acacus Libya 10.50 25.50 2 21 Adrar Bous Niger 20.29 9.01 1,2,3 27,35,36 22 Tin Ouaffadene Niger 20.18 9.19 1,2,4 27,35,36 23 Fachi Niger 18.10 11.30 2 27,60,61

© 2015 Macmillan Publishers Limited. All rights reserved 62 24 Dibella Niger 17.53 13.13 2 25 Kawar Niger 18.60 13.08 2,3 27,60,61 26 Termit Niger 16.00 11.25 2,3,7 36,63 64 27 Jikariya Lake Nigeria 13.31 11.08 6 28 Bougdouma Niger 13.32 11.67 2,3 35,36 65 29 Kaigama Nigeria 13.25 11.56 6 30 Kajemarum Nigeria 13.30 11.02 2,7 27,36,65-67 68 31 Komadugo Nigeria 12.83 12.00 2 32 Bal lake Nigeria 13.31 10.95 6 27,65 33 Wadi Fesh Fesh Sudan 18.90 17.90 2,3,7 27,69-72 34 Bahr el Ghazal Chad 15.42 18.02 2 73,74 35 Yoa Chad 19.03 20.31 2,6 75,76 77-79 36 Farafra Egypt 27.10 28.10 2 37 Dakhla oasis Egypt 25.23 29.42 2 80,81 38 Selima Oasis Sudan 20.92 27.68 2,3 27,71,82-86 39 Dry Selima Sudan 21.37 29.17 1,2,3 27,69,84,87,88 40 Oyo Sudan 19.27 26.18 2,6 27,82,83,89 90 41 Lake Gureinat Sudan 16.97 27.30 1,7 42 Meidob Hills Sudan 15.35 26.35 2,7 27,84,91 43 Atrun Sudan 18.17 26.15 6 27,71,72,83 27,71,72 44 ridge lake T175 Sudan 16.60 27.63 2,3 45 Wadi Howar Sudan 17.60 27.50 2,6 27,71,72,92,93 46 Wadi Mansurab Sudan 15.42 32.22 1,2 94,95 97 47 Rukwa Tanzania 8.42 32.75 3 98 48 Massoko Tanzania 9.33 33.75 3 99 49 Tana Ethiopia 12.00 37.25 5 50 Abhe Ethiopia 11.10 41.40 2,3 100,101 102

51 Abiyata Ethiopia 7.30 38.40 3 Gulf of 103 52 P178-15P Aden 11.96 44.30 5 104 53 NIOP 905 Arabian Sea 50.00 10.00 4 105-107 54 Ziway-Shalla Ethiopia 7.50 38.40 2 108 55 Tilo Ethiopia 7.06 38.09 1 109 56 Kuruyange Burundi 3.83 29.68 6 57 Rusaka Burundi 3.43 29.62 6,7 110,111 112 58 Muchoya Uganda 1.28 29.80 6 113 59 Edward Uganda -0.42 29.58 3,6,7 114 60 Turkana Kenya 2.67 36.50 2,3 61 Victoria Tanzania 1.00 33.00 1,3 115,116 117

62 Mt Satima Mire Kenya 0.30 36.58 7 18 63 Challa Tanzania -3.32 37.70 5 118 64 Tanganyika Tanzania -6.70 29.83 5 119 65 Cheshi Zambia -9.08 29.75 3 66 Tilla Nigeria 10.38 12.13 1,2,3 120,121 67 Segedim Niger 10.17 12.78 6 27,60,122 68 Bosumtwi Ghana 6.50 -1.42 2,5 this paper 123 69 Sélé Benin 7.15 2.43 6 124 70 Dangbo Benin 6.61 2.60 6 124 71 Dogla Benin 6.53 2.37 6 124 72 Yeviedie Benin 6.53 2.38 6

76 Bilanko Congo 15.35 -3.52 6 127-129 77 Barombi Mbo W Africa 4.67 9.40 6,7 130 78 MD03-2707 GOG 2.50 9.39 1,3,7 131-133 79 Njupi Cameroon 6.45 10.32 6 80 Shum Laka Cameroon 5.85 10.05 6 134,135 81 Bafounda Swamp Cameroon 5.53 10.33 6 131,136 82 Mbalang Cameroon 7.32 13.73 6 137,138 83 Ossa Cameroon 3.80 10.75 6,7 139,140 141 84 Nyabessan Cameroon 2.67 10.67 6 142 85 Lake Goualougo Congo 2.16 16.51 6 143 86 Mopo Bai Congo 2.24 16.93 6 144 87 Lake Tele Congo 1.33 17.17 6 88 Lake Maridor Gabon -0.17 9.35 6 131,132,145 89 Lake Nguene Gabon -0.20 10.47 6 131,132,145 90 Lake Kamalete Gabon -0.72 11.77 6 131,132,145 31 91 Congo GOG -5.58 11.22 5 92 Coraf Congo -4.75 11.85 6 131,132,146 93 Kitina Congo -4.25 11.98 6 131,132,147 94 Sinnda Congo -3.84 12.80 6 132,148 95 Songolo Congo -4.76 11.86 6,7 131,132,149 96 Ngamakala Congo -4.07 15.38 6 146,150

Proxy types: 1- oxygen isotopes, 2- stratigraphy/lithology, 3- biological indicators, 4- terrigenous fluxes, 5-

δDwax, 6- pollen, 7- geochemistry